14:00:33 #startmeeting Glance 14:00:35 #topic Agenda 14:00:37 #link https://etherpad.openstack.org/p/glance-team-meeting-agenda 14:00:38 Meeting started Thu Dec 10 14:00:33 2015 UTC and is due to finish in 60 minutes. The chair is flaper87. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:39 o/ 14:00:39 Not a very long agenda, yet. HOW'S EVERYONE DOING? :D :D :D 14:00:39 o/ 14:00:39 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:41 o/ 14:00:41 0/ 14:00:42 The meeting name has been set to 'glance' 14:00:48 o/ 14:01:15 flaper87: still sick, but y'know :P 14:01:19 o/ 14:01:23 let's give some more minutes for folks to join 14:01:30 sigmavirus24: :( get well soon! 14:01:35 btw, happy to have you back :D 14:01:43 I'm happy to be back 14:02:14 sigmavirus24: welcome back! 14:02:26 ok, let's get started. I still have a crappy internet connection so, expect some lag from me :D 14:02:32 tihs time is not my brain's fault 14:02:47 flaper87: sure ;) 14:02:55 :P 14:02:59 #topic Teams Updates 14:03:12 artifacts nikhil ativelkov 14:03:19 o/ 14:03:24 Hi 14:03:35 so, we have two specs on review now 14:03:50 one is for glance v3 -> glare v0.1 14:04:10 ativelkov: I put the v3 -> 0.1 in the specs prio list 14:04:14 another one is a public API description, i.e. the one which addresses the defcore and API-wg concerns 14:04:32 flaper87: I saw that, thanks! 14:04:56 awesome, I think I commented on the other one but was just a quick comment. We'll need API-WG feedback on that one but yeah, lets get it reviewed 14:04:57 Actually, that one was written by mfedosin, but he is on PTO, so I may answer any questions on that 14:04:59 ativelkov: do you have idea about timeline? When would you need the review powar on majority parts of this work? 14:05:43 jokke_: the migration patch is already submitted. it's still WiP, but I expect it to complete today or tomorrow 14:05:57 #chair nikhil 14:05:57 Current chairs: flaper87 nikhil 14:05:59 oh it's on that good shape ... cool 14:05:59 ativelkov: i left come comments on your spec last night 14:06:08 nikhil: in case disappear :D 14:06:10 #link https://review.openstack.org/#/c/255274/ 14:06:19 flaper87: ah ok, sure 14:06:32 rosmaita: thanks! Didn't check it yet, will have some cycles for that later today 14:06:53 from a review pov, right now the focus should be on the migration spec and patches implementing that spec 14:06:56 o/ 14:07:01 not saying we shouldn't review the other one 14:07:05 I just want to quickly sync some major decision points on that migration 14:07:09 ativelkov: nothing major, just some clarifications 14:07:22 shooot 14:07:33 what should be the default glare port? I suggest 8383, but open for other proposals 14:07:42 9393 14:07:48 I guess that's what you meant 14:07:55 oh, yeah, 9393, sure 14:08:17 +1 14:08:19 And what should be the keystone service name and type? I propose type='artifact' and name='glare', but feel free to suggest something else 14:08:22 9393 would be my preference as well 14:08:22 I think jokke_ checked it and it was free 14:08:26 or was it mfedosin 14:08:26 o/ 14:08:34 it was at least during summit still 14:08:40 why not 9293 or 9392 just to be funny? 14:08:53 sorry to bring this up late but do everyone feel ok with name "glare"? 14:09:02 sigmavirus24: because this is serious business and we're not funny here ;) 14:09:02 ooo why not randomly bind to one of those two? 14:09:13 * sigmavirus24 shouldnt' be allowed in meetings 14:09:15 I do feel ok with glare 14:09:19 nikhil: I'm meh 14:09:19 nikhil: i am +1 on Glare 14:09:22 I think as a part of initial phase it seems ok but when you bring up keystone name it sounds a bit weird 14:09:22 I liked it 14:09:27 I don't want glance glaring at everyone but whatever 14:09:38 I would prefer the name "Art" 14:09:46 nikhil: Arti 14:09:49 ativelkov: re service catalog, I think that's fine but please sync with sdague or some keystone folks on that 14:09:55 prolly should send that to ML 14:10:03 sigmavirus24: yeah, that's safer 14:10:06 :_ 14:10:08 :) 14:10:19 nikhil: makes Artifacts sound like your neighbor who is super friend "Hey Arti!" 14:10:29 lol 14:10:30 yeah, nikhil, thanks, ML is a good idea 14:10:31 * sigmavirus24 still wonders why flaper87 invited him back to the core team :P 14:10:32 ok, let's move on 14:10:34 sigmavirus24: but they are! 14:10:43 jokke_: I know! 14:10:46 sigmavirus24: I'm already regrettningit 14:10:48 :P 14:10:49 sigmavirus24: haha, arti in hindi means workship 14:10:58 flaper87: we still have time to block that? 14:11:01 nikhil: all the better :) 14:11:08 :) 14:11:09 jokke_: you already +1'd it :P 14:11:16 ok ok, let's move on 14:11:29 sigmavirus24: I know, but I have friends (believe or not) :P 14:11:30 what else on artifacts? 14:11:45 jokke_: sigmavirus24 friendly notice: shut up! 14:11:47 :P 14:11:49 well, nothing new for now 14:12:02 flaper87: there are tiny things to be aware of 14:12:09 logged here: 14:12:10 nikhil: shoot 14:12:11 #link http://eavesdrop.openstack.org/meetings/glance_artifacts_sub_team/2015/glance_artifacts_sub_team.2015-12-07-14.04.html 14:12:13 done 14:12:21 okidoki, I'll read through 14:12:24 thanks for the heads up 14:12:36 np 14:12:38 Drivers updates: 14:12:58 Let me start with saying that I've added the reviews priorities for specs to the drivers meetting's agenda. To all drivers and people interested in providing feedback, please, take a look here: 14:13:00 #link https://etherpad.openstack.org/p/glance-drivers-meeting-agenda 14:13:02 Also, if you think your spec should be a priority, let me (or any of the drivers) know and we'll evaluate the urgency based on the current load. 14:13:04 Image Import process broken down into sub-tasks (flaper87) 14:13:06 #link https://review.openstack.org/#/c/232371/ 14:13:08 There have been more comments and updates on the spec. If you haven't followed-up on it, this would be a good time to re-read it as I believe we're getting closer to an almost-ready status. 14:13:10 In addition to this, we've broken the spec down into several work items and we'll keep expanding the list as new items/parts are agreed on. This is fairly new and it happened yday, I believe. If you're looking forward to contribute to this work, I'd highly recommend you to go through this list, provide some feedback on the work items and take them :) 14:13:12 Note that there are some dependencies in this work items so please, respect that. 14:13:14 To all reviewers, whenever you're reviewing patches for this work, make sure you have the spec openned right next to the patch you're reviewing :D 14:13:54 questions? concerns? feedback? 14:14:21 can we agree that bikeshed is called bikeshed? 14:14:38 jokke_: it's now called "stage" 14:14:43 jokke_: no, that actually changed in PS8 14:14:45 :D 14:14:54 but we can change it back 14:15:00 ++ 14:15:01 rosmaita: sshhh 14:15:02 if there's popular demand 14:15:06 sorry 14:15:06 don't you know sigmavirus24 is back ? 14:15:13 don't get him started 14:15:19 stage sounds good to me 14:15:20 sigmavirus24 wants to call it the "yak pen" 14:15:31 but not sure if it's a noun or verb 14:15:32 lol 14:15:38 nikhil: exactly! 14:15:45 it's both 14:15:47 :) 14:15:48 * sigmavirus24 is being quiet 14:15:50 thought so 14:15:51 BAM! mindblown 14:15:52 then we can call the working dir in the worker code "backstage" ... sounds good to me :) 14:16:06 jokke_: LOOOOOOL, love that 14:16:06 green room? 14:16:08 hahahahhaha 14:16:12 are tasks stage hands now? 14:16:16 ok ok, let's move on 14:16:25 Cross project updates 14:16:29 nikhil: floor is yours 14:16:38 So, that one got feedback 14:16:40 so many updates, this is good 14:16:42 and there are two things 14:16:52 it's hard to get the lib on list of exceptions 14:16:57 lib == glance_store 14:17:06 as things are we will see when we hit a situation 14:17:18 as the process is sorta complexly variable 14:17:31 but overall I like the idea of governance tag 14:17:39 and about to add +1 to that idea 14:18:03 overall the spec ( https://review.openstack.org/#/c/226157 ) is in good shape 14:18:40 mmhh, but that's not a tag, though. it's a x-proj spec. The tag exists already and it's called assert:follows-standard-deprecation 14:18:45 * nikhil done. was thinking wrongly, spec was linked. 14:18:49 unless they are talking about a new tag and I missed that part 14:18:59 nikhil: thanks for the updated 14:19:02 no I think flaper87 you are correct 14:19:06 I don't know which tag they are talking about 14:19:12 it's a comment from doug 14:19:30 ah, I probably missed that comment 14:19:34 I think that plan will be reverted next cycle again (branchless clients and libs) because all the hell will break loose but lets see 14:19:35 I'll read through 14:19:51 ok, let's move on 14:20:07 #topic New home for release schedule (flaper87) 14:20:09 Just a heads up that the release schedule home has changed. It's not longer in the wiki page but in the openstack docs url. Here's the link: 14:20:11 #link http://docs.openstack.org/releases/schedules/mitaka.html 14:20:13 I'll be updating that schedule to add Glance's timelines and schedule activities. 14:20:49 questions? feedback? Just wanted to make sure you were aware of this 14:21:12 * flaper87 wants a per-project .ics with scheduled activities just like we have fro meetings 14:21:20 * flaper87 will propose that 14:21:26 ++ 14:21:43 ok, moving on 14:21:45 that would be super handy 14:21:54 #topic glance_store's drivers deprecation (flaper87) 14:21:56 #link http://lists.openstack.org/pipermail/openstack-dev/2015-December/081966.html 14:21:58 I sent that email out a couple of minutes ago. I don't expect folks to have read it already but I'd like to encourage everyone to do so. Feedback would be awesome as that's going to make some people sad/mad and it'd be great to have the community on the same page so that we can provide the best feedback to consumers of the library and, of course, stand behind on the decisions made. 14:22:31 I noticed swift wasn't in the list of glance-maint 14:22:56 nikhil: I knew you would be one of them :D 14:23:00 and I just found out the explanation 14:23:03 ;) 14:23:17 hope the reasoning makes sense but I'm happy to discuss that further 14:23:18 so the plan is to run all glance functional tests against master glance_store? 14:23:26 (I think that's a good idea) 14:23:32 yeah 14:23:43 master & stable!? 14:23:45 pretty much the same way cinder does w/ third-party drivers 14:23:50 nikhil: yup 14:23:58 flaper87: sounds good 14:24:46 happy to discuss this now if folks have read the email already or we can add it back to next week's meeting due to the short notice 14:24:56 or both 14:24:59 :P 14:25:22 next meeting sounds good to me to be able to ponder on things 14:25:29 ++ 14:25:31 ok, I'll add it there 14:25:35 wow, that's some email! 14:25:38 I'm good with the rest of it 14:25:45 yeah, I thought so too 14:25:53 what happens if we don't get people to volunteer? 14:26:01 okidoki, let's move on :D 14:26:12 kragniz: the email says, we'll remove that driver 14:26:15 me too, looks good and wasn't a surprise since we have been talking about 3rd party CI for sometime 14:26:24 kragniz: we will! or the drivers go in a cycle or two :) 14:26:29 I'll try asking Mirantis's ceph team if they want to volunteer maintaining the ceph driver 14:26:40 cool, in O 14:26:42 ativelkov: sounds awesome! 14:26:43 kragniz: passive reach out to get more open source upstream friendly 14:27:10 more people* 14:27:14 ok, let's move one 14:27:23 #topic Release notes 14:27:27 jokke_: was that you? 14:27:30 yeah 14:27:49 floor is yours 14:27:52 so quick heads up for people. We moved to utilize reno for release notes 14:28:01 w0000000h000000 14:28:04 #link http://docs.openstack.org/developer/reno/ 14:28:36 all our projects have now moved, correct? 14:28:42 (store + client) 14:28:42 tons of small files ahead so please look into the documentation and prepare yourselves that Liberty and Mitaka+ will require those 14:28:52 I think yes, glance_store was the last one to join 14:28:56 cool 14:29:06 so lets start including those release notes into the changes, please 14:29:30 so searchlight team added some good guidelines for when and how to add notes 14:29:32 #link https://github.com/openstack/searchlight/blob/master/doc/source/feature-requests-bugs.rst#release-notes 14:29:36 jokke_:anything reviewers should keep in mind ? 14:29:50 nikhil: ++ thanks 14:29:55 we should probably have that as well 14:29:59 in the contirbuting docs 14:30:10 besides there's this: 14:30:15 #link http://docs.openstack.org/project-team-guide/release-management.html#how-to-add-new-release-notes 14:30:17 shouldn't these be somewhere common and link to that? 14:30:19 contributing* 14:30:28 flaper87: nothing special comes up to my mind atm. good idea ofc check the output of the rel notes job to verify that they actyally get included 14:30:33 kragniz: that was project specific 14:30:41 openstack specific is the link above 14:31:07 awesome stuff! glad to see the rel-notes process improved 14:31:09 other than that lets get that going so we can later on agree requiring those and providing - votes for not including them 14:31:46 ++ 14:31:51 one thing to remember is who these notes are for: end users and deployers 14:32:03 sounds good to me! Any questions? objections? 14:32:05 that's it from me 14:32:22 kragniz: ++ 14:32:25 more descriptive the better (explicit vs implicit) 14:32:26 extra noise in release notes makes things harder for them to pick out important things 14:32:41 kragniz: specially service & client 14:32:52 yeah, but good meta-info clears that confusion too! 14:32:59 store is bit there and there ... they are more for us ;) 14:33:02 (for example, end users don't care about us refactoring tests) 14:33:03 * flaper87 likes that kragniz always brings that up as a nice reminder 14:33:21 flaper87: if I say it enough times! 14:33:46 kragniz: jokke_ we should make a gate check for those release notes 14:34:02 kragniz: re tests, that's debeatable. Some consumers of these libraries use the tests shipped in the library 14:34:05 We should automate that kind of stuff instead of relying on reviewers to remember to -1 things 14:34:07 sigmavirus24: what would it check? 14:34:16 kragniz: but deployer does care if we refactor the test to validate that it still tests their use-case 14:34:16 kragniz: that the commit includes release notes? 14:34:46 but I agree "fixed a typo" is not useful 14:35:25 sigmavirus24: let's get the people used to that first and then we can see how we gate it fitering those typofixes etc. out 14:35:40 sigmavirus24: do you mean for all commits? 14:35:46 flaper87: okay, I guess so 14:35:50 kragniz: gate checks can be skipped 14:36:03 It would probably take some tuning but we could certainly do it 14:36:49 (unless the typo is in a function name) 14:37:03 as said lets think that _after_ we get the community used to this (and everyone having at least their first rel note merged) 14:37:09 sigmavirus24: it's an interesting idea 14:37:21 but yeah, we should get used to the process first 14:37:39 anything else on this topic? 14:37:44 that's all from me 14:37:57 ditto 14:38:07 coolio 14:38:13 #topic Open Discussion 14:38:24 nikhil: ? 14:38:36 so this one 14:38:47 #link https://review.openstack.org/#/c/243348 14:38:54 cli stuff, right? 14:39:09 yes, sorry. the title is: 14:39:10 Deprecate individual CLIs in favour of OSC 14:39:26 It's gaining good traction 14:39:30 lots of +2s 14:39:34 it'd be great totget people's feedback on that spec 14:39:35 and I overall like the idea 14:39:44 * flaper87 +2'd it and jokke_ is going to hate him for that 14:39:44 I have been thinking about one thing though 14:39:55 thanks nikhil 14:40:05 flaper87: that's your call ... will -1 it myself ;) 14:40:21 I'm +1 on that 14:40:29 it's a pain to use all the different clis 14:40:31 nikhil: what's that? 14:40:33 besides the import stuff 14:40:39 the location management stuff 14:41:07 I find it hard to believe it's out there on the stage with configuration etc at the glance backend 14:41:24 I don't think anything will happen until N (or even O) but t's good to keep in mind that's likely going to happen 14:41:33 the UX is pretty bad and am not sure if it's interop fussy topic 14:42:25 nikhil: that's related to OSC, right? 14:42:29 anyways, it just seems like a bad idea to port that to OSC 14:42:33 yes 14:42:37 it'd be great to bring all this feedback to the m-l and the OSC team 14:42:51 as said before I'm happy to spin off glancecli project forking the last version of cli we ship if we're going to this direction :P 14:43:01 I have always been of the opinion that location management is for the case where operators trust their users (private cloud) 14:43:05 nikhil: I'm curious to know how all that will work on OSC 14:43:29 exactly 14:43:41 jokke_: you're welcome to do that :P 14:43:49 so, I am not sure at this point if we can remove the glanceclient CLI 100% 14:43:50 cool, do let me know when you have some extra feedback 14:44:00 erm, some answers for that 14:44:07 that was a hell of a freaudian slip 14:44:29 nikhil: I don't think people want to remove it right now 14:44:30 nikhil: perhaps a new thread on the m-l about this? 14:44:46 kragniz: I think nikhil meant "entirely and not when that will happen 14:44:46 nikhil: more tell people that the recommended thing is OSC 14:44:46 ok, noting it down 14:45:13 flaper87: right, that's what I meant 14:45:21 the spec says it'll stay for 2 more cycles after it's marked as deprecated 14:45:37 flaper87: okay 14:45:58 * kragniz only skimmed the spec, sorry :( 14:46:02 ok, one more heads up from me: The spec freeze date is approaching so, spec writeers, please, I beg you ,address comments in a timely manner so we can work with you on getting those merged 14:46:49 There have been reviews to specs and quite some lag on addressing those feedbacks 14:46:52 well the spec proposes clearly 2 cycles to remove native clients 14:47:01 I've mentioned this in the email I sent to the m-l last night 14:47:27 jokke_: you have 2 cycles to fork the CLI 14:47:57 anything else? anyone? 14:48:34 I want to propose a video conference for artifacts. 14:48:58 nikhil: to demo it or to talk about the new API ? 14:48:58 sometime before the holidays (depending on your locality) 14:49:07 * sigmavirus24 has to cut out early 14:49:17 and after 16th dec (when Mike is back and settled hopefully) 14:49:26 Sounds good to me 14:49:33 flaper87: no, to increase momentum and get early feedback 14:49:35 Mike will be back on Monday, I think 14:50:01 anything that we need to talk about it actually 14:50:06 I'm happy to join and schedule it. WE can use my bluejeans bridge as we did for the import process 14:50:13 I can propose some agenda and send out etherpad 14:50:17 nikhil: I like the idea. Need to try to invite the guys from community app catalog 14:50:25 ativelkov: +1 14:50:35 nikhil: please do, I'll setup the meeting on bluejeans so we can record it and what not 14:50:54 oh yeah ... FYI I'll be out of picture between Tue 15th and Tue 5th of Jan 14:50:58 great 14:51:10 oh noes 14:51:19 flaper87: I will sync offline when I have something ready 14:51:21 so don't expect reviews, actions, e-mails, replies etc. from me during that time :D 14:52:09 jokke_ please, enjoy your holidays. If we see an email, review, rant, joke from you on that period of time, we'll revoke your -2 rights and just leave you with +1 and +2 14:52:22 I hope we have another one later in the cycle when things are more through the gate and jokke_ will be there then 14:52:26 actually, just +1 14:52:33 and no +2 14:52:36 :P 14:52:40 +A* 14:52:42 flaper87: ninja reverts ... you won't even notice them :P 14:52:53 jokke_: +1 on enjoying your holidays 14:53:17 flaper87: I promise to stay more within the Holiday scope than you did 14:53:21 :) 14:53:30 touche 14:53:39 damnit, ppl always use that against me 14:54:05 ok,if there's nothing else, we can call it a meeting 14:54:29 thanks everyone! see y'all next week. 14:54:33 thanks all! 14:54:35 #endmeeting